Gemini AI Power Up What ChatGPT Is Missing
Google's I/O 2025 keynote was a significant showcase for Gemini AI, presenting a suite of features that aim to set it apart from competitors like OpenAI's ChatGPT. While last year's I/O 2024 saw OpenAI strategically counter-program Google, this year, Google's presentation of Gemini's enhanced power, speed, and deeper integration into its ecosystem stood largely unchallenged.
The author, a regular user of ChatGPT, expressed envy towards several new Gemini functionalities, particularly those leveraging Google's vast suite of apps and services—an advantage OpenAI cannot easily replicate. Even though some features are still in development, they signal Google's ambition to take a substantial lead in the AI race.
Here are eight standout Gemini features highlighted at I/O 2025 that present a challenge to ChatGPT:
Deep Think for Advanced Reasoning
Building on existing deep research capabilities, Google introduced Gemini Deep Think. This enhanced reasoning mode is designed to provide significantly better and more nuanced responses. While still in safety evaluation, Deep Think promises a leap in AI's analytical abilities.
Gemini 2.5 Pro Deep Think feature tops benchmarks. Image source: Google
Versatile AI Agents Across Google Products
Google is significantly upgrading its AI agent capabilities, exemplified by improvements to Project Mariner. These AI agents are set to be integrated into various Google products. For instance, Google Search's AI Mode will be able to monitor product price changes and even make purchases on the user's behalf. This marks a significant step towards AI that actively assists in daily tasks.
Gemini Live and Project Astra: Your Proactive AI Assistant
Gemini Live is evolving into a sophisticated AI assistant, as demonstrated with Project Astra. It's not just about responding to queries; it's about proactive, agentic behavior. This includes making calls, placing online orders, browsing the web for information, and even understanding a multi-person conversation while staying on task. Gemini Live can also pull information from other Google apps and remember user history for context.
Watch: Project Astra | Exploring the Capabilities of a Universal AI Assistant
While many advanced features are forthcoming, current camera and screen-sharing capabilities are already available on Android and iPhone. This contrasts with OpenAI's Operator, which remains limited to a high-cost tier.
Revolutionizing Online Shopping with AI Mode in Search
Google Search's AI Mode is getting powerful shopping features. Beyond conversational product searches, it will introduce a "Buy for me" function, allowing the AI to monitor prices and purchase items when they drop. A standout feature is "Try it on," where users can upload a photo to see how clothes might look on them, leveraging a special AI model.
AI Mode will also assist with finding and buying event tickets, making restaurant reservations, and scheduling appointments—tasks currently beyond ChatGPT's direct capabilities.
Unparalleled Personalization Across Google Services
Gemini's deep integration with Google's ecosystem allows for a high degree of personalization, albeit with privacy considerations. Demonstrations showed Gemini drafting emails in Gmail using the user's writing style and pulling relevant information from documents or photos. Gemini Live can access Gmail data for context and remember details like a user's pet for tailored search results. AI Mode can also offer personalized suggestions based on search history and connect to Gmail for richer context. ChatGPT lacks this native app integration.
Gemini: Gmail email personalization based on the user’s writing style. Image source: Google
Gemini: Gmail email personalization based on email history. Image source: Google
Flow: Next-Generation AI Video Creation
Google's Flow was a major highlight, presented as a potential step beyond OpenAI's Sora. Flow enables the generation of videos complete with audio, maintaining character and scene consistency. Users can also continue editing these projects outside the AI program. While OpenAI has Sora, Flow's described capabilities appear advanced.
Watch: Introducing Flow | Google’s New AI Filmmaking Tool
Seamless Real-Time Translation in Video Calls
Gemini is bringing AI-powered real-time translation to video-chatting apps like Google Meet. While real-time translation isn't new, its integration into widely used video call platforms by Google is a significant step that ChatGPT currently doesn't offer directly within a similar communication product.
The Dawn of Gemini-Powered Hardware
Google provided public demos of its AR/AI and AI-only smart glasses, the first devices developed with Gemini at their core. Despite some demo glitches, these AI wearables showed functional promise. Android XR glasses are expected later this year, potentially offering an ideal platform for AI interaction, enhanced by Gemini's unique capabilities. This is an area where OpenAI currently lacks first-party hardware, giving Google a head start in AI-integrated wearables.
The advancements showcased at I/O 2025 suggest Google is aggressively pushing Gemini's capabilities, creating a compelling challenge for OpenAI and setting a new bar for AI assistants. The focus on deep integration, real-world task execution, and dedicated hardware could reshape how users interact with AI.
